Title: The Mechanic and the Luddite: A Ruthless Criticism of Technology and Capitalism

Author: Jathan Sadowski

Completed: July 2025 (Full list of books)

Overview: I’ve been listening to This Machine Kills podcast and enjoy their take on current technology developments enough that when I heard one of the hosts had a new book coming out, I immediately reserved it at the library. This book got deeper into politics and techno-capitalism than most discussions I get into as seen by the copious notes below.

There is a lot packed into this book, but one recurring theme that I feel is often overlooked is that the future technologists are building (and selling) isn’t the only possible future. So often we’re told when the latest tech is released that eventually everyone will be using it (smart phones, the internet, cars, if we go back further). The problem is the people telling us everyone will use the latest tech are also the people selling it. They are not predicting the future, they are marketing it and often wrong (VR/Metaverse, Google Glass, Web3, blockchain). When it comes to new technology, having a Luddite mindset of only accepting new tools that measurably improve our lives is important

Highlights:

• Over time, the financial interest in profit transformed into the social imperative of profit-making. Milton Friedman, Nobel Prize–winning economist, champion of the free market, and granddaddy of neoliberalism explicitly advocated for this transformation. He did not originate this argument, but he
• crystallized it and slapped his name on it in a 1970 essay for the New York Times titled A Friedman Doctrine—The Social Responsibility of Business Is to Increase Its Profits.⁹ The article itself is a convoluted argument about how, actually, the singular pursuit of profit in a system of free enterprise is the most efficient and most ethical way to organize society. At this time, people still mostly thought of finance as only one part of society and believed that corporations had social obligations beyond profit. Today, Friedman’s doctrine is simply treated as common sense. No longer does capitalism need to justify its existence or offer defensive cases for profit-making. The system is now focused on advancing offensive tactics for profit-taking and bulldozing any barriers to its endless expansion.
• Capitalism is also built on the alchemy of abstraction. By this I mean it is a system that excels at taking a concrete, specific thing like the house you live in and turning it into an abstract, universal category called an asset. Or, taking a collection of similar but different things like varieties of apples grown in different places and turning them into a singular, standardized category called a commodity. Even the most basic features of capitalism, concepts like property and wages, which don’t seem strange at all because they feel like they have always been part of society, are ultimately ways capitalism has abstracted how we relate to ourselves, other people, objects, labor, and value
• familiar demons like racism, sexism, ableism, homophobia, transphobia, among others. However, these are not different heads of a hydra acting independently from each other and only occasionally working together to cause havoc. They are coordinated parts of a collective whole. Capitalism needs to construct hierarchies of power based on social differences and to further inflame existing forms of control and inequality. It creates ways of dividing people and devaluing human life, making people easier to exploit, oppress, and discard like used up resources.
• If subsumption continues, then these practices are totally remade and reorganized according to the imperatives of capital such that they become inherently capitalist in nature. We can see this shift historically when early capitalists initially captured profit from forms of artisan labor that existed before capitalism by enforcing new property regimes on their work—for example, by making artisans rent the tools needed to craft goods. Then with the rise of industrial capitalism, these forms of labor were fully transformed by the factory system; they were absorbed into capitalism, becoming appendages of capital. Subsumption has now come for everything.
• Similar processes of subsumption are also enacted by technological systems. First the application of, for example, some new smart device is used to augment an already existing process. Before long, the process itself is being changed to feed the needs of smart tech: data is collected constantly about every aspect of its use and its functions are controlled automatically. So when a coffee maker is made smarter it gets some new features that usually add a bit of convenience and connectivity. Oh look, I can turn on the coffee maker using an app or I can get real-time updates on my bean levels. Kind of weird, but also maybe cool and useful. But then, very quickly, the coffee maker gets an update and now it’s sharing all the data it collects about when, how, and what kind of coffee you drink—plus maybe other stranger data like pictures of your kitchen—with third parties who then use it for their own (unknown) reasons. And the coffee maker keeps sending notifications with sponsored advertisements for coffee beans and insisting that I set up an account to automatically order beans when it senses my supply is low. Also the coffee maker won’t work unless it is always connected to the internet so it can stay in constant contact with the manufacturer’s servers in another country.
• when digital technology enters into a sector it does so not only by upgrading their capabilities and selling them products but also by trying to transform, disrupt, colonize, and monetize that sector. For example, cars are now described as complex computers on wheels that produce enormous data about their operations (and drivers). This is why the recent microchip shortage was, for many people, felt most strongly in the car market. Cars now run on microchips, and there just weren’t enough to go around, which caused scarcity and rising prices. Back in 2018, the CEO of Ford declared that the future of automotive manufacturers is in being data miners. [Ford] could make a fortune monetizing data. They won’t need engineers, factories or dealers to do it. It’s almost pure profit
• As historian Leo Marx has shown, the idea that improved technology means progress is an old one that traces a long way back to the Enlightenment project of using scientific knowledge and technological power to dominate nature and accumulate wealth. ³⁰ This ideology evolved to prioritize technocratic progress on the assumption that other forms of progress (e.g., social, economic, political, moral) and other values (e.g., freedom, justice, autonomy, equity) will follow as long as the technocratic machine continues to advance. (Now, however, our expectations of how technology can improve our lives have been lowered to the degree that innovation doesn’t even need to be improved in any meaningful sense to be lauded as progress
• Beer put forth a maxim that I think should be an essential rule of thumb for analyzing technology: The purpose of a system is what it does. There is after all, no point in claiming that the purpose of a system is to do what it constantly fails to do.
• these claims just propel cycles of wild hype where a stupendous amount of money and noise results, at best, in products with dubious value. Here I’m thinking of the blinding flash of Web3, NFTs, crypto, and so on. ² Many investors and consumers lose their shirts, a few make out like bandits, and some from both sides go do the same exact thing with another venture in the next hype cycle while calling themselves serial entrepreneurs who know the value of failing fast.
• The mechanic knows how a machine is put together, how its parts function, and what work it does. The Luddite knows why the machine was built, whose purposes it serves, and when it should be seized—in both senses of stopped or taken, destroyed or expropriated. We should always strive to embody both of these models. Neither is sufficient on its own.
• In a literal sense, being a mechanic means having a trade, a profession, a certification, a degree, a set of skills. But in the analytical sense I’m using here, being a mechanic is as simple as pursuing a curiosity about how the world really works and what you can do in it.
• Luddism is not a naive belief but a considered position. It does not come from being ignorant of technology but from being informed of its functions. Once you know what Luddism actually stands for, I bet you will begin identifying as a Luddite too—or at least be more sympathetic to the position than you might have thought.
• The first step to being a Luddite is simply opening yourself to the radical possibility of evaluating the technologies that fill our lives and determining whose goals they primarily advance. Not all innovations deserve to exist, and many should never have been created in the first place. Yet we assume their legitimacy and acquiesce to their existence merely because they have already been made. Silicon Valley demands we accept their products like a cargo cult receiving gifts from the gods. Instead we need an approach modeled more closely after Marie Kondo. For every technology, we should hold it up and ask—not, Does this thing spark joy?—but, Does this thing contribute to human well-being or social welfare? If not, toss it away!
• The mechanic knows how a machine operates, how it is put together, and how it can be repaired or reengineered. The Luddite knows why the machine was built, whose purposes it serves, and when it should be disassembled or destroyed. By becoming mechanics and Luddites, we get to the heart of how these systems work, who they work for, and what we can do to change them. Together these models provide us with the tools necessary for rejecting the systems thrust on us by others and, in their place, making our own future.
• The mark of venture capital can be found on the semiconductors and mainframe computing of the 1960s and 1970s, the personal computing and software apps of the 1980s, the internet and e-commerce of the 1990s, the mobile and social web of the 2000s, the gig economy and x-as-a-service platforms of the 2010s, and the crypto assets and generative AI of the 2020s. I should also note that VCs have been active in the life sciences and biotech start-ups since the 1980s, financing ventures in areas like pharmaceuticals, genetics, and medical devices—although investment has been on a steady decline for the last twenty years ¹⁰ and has shifted toward shorter-term, lower-risk, higher-return investments ¹¹ like developing hardware and software for medical applications. (Theranos was meant to be the crown jewel of Silicon Valley biotech.)
• VCs choose technologies that fit the economic conditions they need to prosper. “A vicious circle emerges,” Cooiman writes, “allowing these few [investors] to collect more capital, be more attractive for high-potential start-ups, and, with their network and experience, effectively not only pick but create winners, which again increases returns and overall attractiveness.”
• they make choices about what kinds of technologies count as “high potential” and which pathways are not worth pursuing. ⁴⁰ A technology transfer officer and former venture capitalist told Lee: “There’s just countless examples of that, where poor quality innovation is what actually makes it to market, because of the team, the network, the location, the hype, the everything.” The reason why so much investment is sunk into software is not because that’s where the most or best innovation necessarily happens. It is because, generally, software is cheaper to build, quick and easy to scale, able to serve a large market, and lower risk compared to many other types of technology.
• All the world’s a casino, and all the VCs and entrepreneurs merely players; they have their exits and their entrances.
• the reason why ubiquitous surveillance is built into our digital society is not because it’s a technical requirement or inevitable feature but because it’s valuable for capital.
• The sudden blockbuster success of OpenAI’s consumer technologies in late 2022 sparked an enormous focus on (generative) AI that now defines the tech industry. This was also the same exact time when the Web3/crypto economy was in spectacular collapse. Many tech companies and venture capitalists were searching for the next big thing to hang their promises for the future on. To underscore this point, a prominent venture capitalist, Jason Calacanis, tweeted in June 2023: “If you’re in crypto pivot to AI.”
• The rising costs of AI would be much higher if they also accounted for a major subsidy underlying these companies. The necessary training data was largely free, or at least deeply discounted, by a data collection practice more commonly known in other sectors as theft.
• synthetic data is “generated by algorithms and for algorithms,” which means it can be a much easier, cheaper alternative to real data. ⁶⁵ While synthetic data can be technically better for some applications, its use should still require the kind of stringent human oversight and verification that companies are trying to eliminate. Sam Altman has claimed “that soon all data will be synthetic data,” in large part because human-created data is comparatively expensive and fresh sources of the real stuff are becoming more scarce and harder to tap.
• Whether it’s content moderation for social media or facial recognition for police surveillance, ³ claims about the capabilities of AI systems are more abundant and incredible than ever before. While we are led to believe that these smart technologies are solely powered by neural networks, as tons of research and reporting have shown, much of the cognitive labor essential to their operations likely comes from an office building full of (low-waged) workers in popular outsourcing destinations like the Philippines or India or Kenya.
• Considering what we know about the data used to train this AI system—scraped from the toxic waste dump of the internet—cleaning those inputs and creating guardrails for ChatGPT’s outputs is a complex problem. OpenAI’s solution was to build an “additional AI-powered safety mechanism” that would ensure the proper safeguards were in place for the consumer-facing chatbot. And how was that backend AI created? As a Time investigation revealed, Kenyan workers making less than two dollars per hour had to label the training data by going through “tens of thousands of snippets of text,” some of it describing “situations in graphic detail like child sexual abuse, bestiality, murder, suicide, torture, self-harm, and incest.” ¹⁸ Only then could the AI safety tool “learn to detect those forms of toxicity in the wild.”
• the on-demand gig platforms like Uber, Instacart, and TaskRabbit should be reframed as “servant apps.” ³⁵ In addition to entrenching a techno-capitalist regime “marked by extreme exploitation and despotic control of labor,” these companies’ value proposition for consumers is based on democratizing access to servants by allowing people to summon workers at their command.
• crucial for understanding technology in capitalism is the quixotic quest by capital to build what I call the “perpetual value machine.” In short, the machine would be a way to create and capture an infinite amount of surplus value without needing any labor to produce that value. Capital has been pursuing this quest for hundreds of years—continually investing, innovating, hyping, failing, and trying again to reach this ultimate goal. Why? Because in addition to finally satisfying capital’s endless hunger for profit, the perpetual value machine would also eliminate the thing that is both the mortal enemy of capital and a vital necessity for capital: the power of labor.
• We tend to think of manufacturing, especially in the automotive industry, as being among the first and most automated sectors of the economy. However, the material reality is far more complex, with most jobs still being done by humans and with technological advancements moving in multiple directions. Benavav cites research by labor sociologist Martin Krzywdzinski showing that Toyota—considered the most efficient car company in the world—has actually been removing robots from its assembly lines to take greater advantage of the flexibility and responsiveness of human workers. ⁴⁹ Less than 10 percent of the assembly work is automated at Toyota.
• At the heart of Silicon Valley lies a dirty secret. Contrary to the popular mythology that it is the greatest engine of value creation—a mythology that is still deeply entrenched, even in the face of recent public skepticism—much of Silicon Valley’s real wealth comes from capturing value. Despite its reputation as the land of innovators, the tech sector is filled with landlords. (Granted they have come up with some very innovative ways to extract rent at scales rarely seen before.)
• Silicon Valley’s primary business model—often called “x-as-a-service”—is based on acting like landlords and treating us like tenants. Notice that tech companies never sell us anything in the sense of transferring ownership. They only offer access to services in exchange for personal data, charging for subscriptions, or paying per usage. That access is then governed by terms and conditions agreements, a type of legal contract that nobody ever reads because they are not meant to be read. Importantly, this model often takes shape as the digital platforms that have become significant forms of infrastructure in society.
• when you buy a smart thing, you only own the physical object; the digital software is licensed—which means leased or rented. This gives the license holder continual access to the object. That access then grants powers like remote control over the object and data collection from the object (and the people, animals, environments, and other devices it interacts with). In effect, by integrating everything into smart systems, companies are able to enact a form of micro-enclosure in which they retain ownership over the digital parts of physical things, and all the rights and powers that ownership entails, even after you purchase it. The deeper we incorporate these smart things into our lives—to the point that they become attached to us and we become dependent on them—the more powerful, invisible, and valuable the corporate control of technology becomes.
• even after spending $50,000 on a family sedan, or $250,000 on a farm tractor, you own a big hunk of metal, rubber, and silicon—but you are only renting the software needed to actually operate the vehicle.
• In her book Undoing the Demos, Brown identifies the 1980s as a key moment in the rise of governance. It was at this point that neoliberal approaches to government were taking effect in the US with Ronald Reagan and in the UK with Margaret Thatcher and were quickly spreading to become a global hegemonic paradigm. This paradigm explicitly aimed “to transfer private-sector management methods to public services and to employ economic techniques such as incentivization, entrepreneurialism, outsourcing, and competition for public goods and services.” ¹² In other words, the point was to make governments operate more like corporations, to replace political issues with financial logics. This was done by importing the models, metrics, values, concepts, and tactics of the FIRE sector into the sociopolitical sphere—with the assessment and management of “risk” being central to these practices of power/knowledge.
• Yet the magnitude of recent climatic disasters is now forcing investors and insurers to confront the fact that they have severely mismanaged the costs and certainty of these catastrophes. So after shooting itself in the kneecap, the FIRE sector then turns the gun on the public and demands they help fix this very dire situation before everybody gets blown away. This is a form of governance that transfers any risks onto public balance sheets, while ensuring rewards are funneled into private coffers.
• Consumers are then targeted with personalized prices that reflect how much people are willing to pay, rather than prices based only on how risky they are compared to other similar people. Emerging models of insurance—with names like “on-demand” or “insurance-as-a-service”—are engaged in trying to make the industry even more dynamic so that insurers can, for example, change prices and policy conditions as often as they want. ⁵³ The aim is to innovate around regulations that restrict how often such changes can occur. Those changes might reflect the dynamic nature of risk, but we should also expect them to optimize for profit.
• It would feel strange to conclude a ruthless criticism of complex systems with an infomercial for the one weird trick to undermine capitalism and reclaim technology. The bourgeoisie hate it! The state cannot stand it! Call now! This book is not the last word.
• Real utopias are found in dynamic mechanisms of social coordination that always secure the changing needs of everybody and meet the changing desires of new generations.